Forecasting The Stock Market With The Hidden Markov Model Based On Investment Sentiment

Studies of behavioral finance have shown that the sentiment has a significant impact on investors decision-making.When feeling positive emotions,investors tend to overestimate investment opportunities and underestimate investment risks,and thus they trade stocks frequently.When feeling negative emotions,investors behave conversely.This investment behavior eventually led to the stock market fluctuating abnormally,and is not conducive to the stable and steady development of the capital markets.Therefore,if the sentiment could be applied,there are new ideas to study the ups and downs of the stock market.This paper intends to quantify investor sentiment,and then build the hidden Markov model to forecast the stock market.Firstly,this paper uses the sentiment-financial decision model to analyze the two paths of investor sentiment affecting the stock market: 'sentiment-cognition-behavior' path and 'sentiment-behavior' path.And then,this paper analyzes how to use the hidden Markov model to forecast the stock market based on investor sentiment.In empirical analysis,this paper first applies the partial least squares(PLS)to construct a comprehensive investor sentiment index.This index presents that Consumer Confidence Index,New Investors Amount,Trading Volume and Turnover rate are positively related to investor sentiment,and the discount rate of closed-end funds is negatively related to investor sentiment.Then,given that China's stock market is susceptible to the policy,this paper studies nine good and nine bad policy events.The results show that the influence of the policy on the stock market has been reflected in investor sentiment,so there is no longer need to consider the policy when applying investor sentiment to predict the stock market.no longer need to consider the impact of policy factors.After that,according to the comprehensive investor sentiment index from 2007.01 to 2015.12,this paper calculates key parameters of the hidden Markov model,and predicts the stock market based on the forecasting principle using MATLAB software.The results show that the longer the predicted time period,the smaller the probability of the outcome.Furthermore,in order to highlight the role of investor sentiment,this paper also uses the Markov chain to predict the stock market,and tests and compares results of these two kinds of two models.The analysis shows that the result of the Hidden Markov model is more accurate.That is to say investor sentiment is effective inpredicting the stock market.